Anti-KC4 humanized monoclonal antibody and cells that express the antibody

Abstract

An anti-KC-4 humanized monoclonal antibody that comprises the variable regions of the light and heavy chains of the anti-KC-4 murine antibody, wherein the light chain has 7 amino acids and the heavy chain has 12 amino acids of the framework regions substituted with amino acid present in equivalent positions in antibodies of a species other than munne, and the constant regions of a human antibody. The antibody may be labeled and/or glycosylated, and is presented as a composition with a carrier. The anti-KC-4 monoclonal antibody is used in diagnostic kits for cancer and in in vivo methods of imaging and treating a primary or metastasized cancer, and in vitro diagnosis and ex vivo purging neoplastic cells from a biological fluid. RNAs and DNAs encode the monoclonal antibody, and a hybrid vector carrying the nucleotides and transfected cells express the peptides.

Description

BACKGROUND OF THE INVENTION

1. Field of the Invention

The invention relates to the in vito and in vivo diagnosis and therapy of carcinomas by means of a specifically targeted humanized mouse monoclonal antibody selectively binding the human KC-4 antigen. The humanized anti-KC-4 mouse antibody comprises the complementarity determining regions (CDRs) of the variable regions of the mouse antibody of the same specificity, and its framework regions having specific amino acids replaced in a predetermined manner, and the constant regions of a human antibody. The humanized anti-KC-4 mouse antibody of this invention is expected to elicit a lesser immunological response in humans than the whole mouse antibody and is therefore considered suitable for in vivo administration to humans. Polynucleotide segments encoding the humanized antibody, a hybrid vector and a transfected host cell carrying the DNA segments encoding the antibody are useful for preparing the peptides disclosed herein.

2. Description of the Background

Carcinomas result from the carcinogenic transformation of cells of different epithelia. Two of the most damaging characteristics of carcinomas are their uncontrolled growth and their ability to create metastases in distant sites of the host, particularly a human host. It is usually these distant metastases that cause serious consequences to the host, since frequently the primary carcinoma may be, in most cases, removed by surgery. The treatment of metastatic carcinomas, that are seldom removable, depends on irradiation therapy and systemic therapies of different natures. The systemic therapies currently include, but not fully comprise, chemotherapy, radiation, hormone therapy, different immunity-boosting medicines and procedures, hyperthermia and systemic monoclonal antibody treatment. The latter can be labeled with radioactive elements, immunotoxins and chemotherapeutic drugs.

Radioactively labeled monoclonal antibodies were initially used with success in lymphomas and leukemia, and recently in some carcinomas. The concept underlying the use of labeled antibodies is that the labeled antibody will specifically seek and bind to the carcinoma and, the radioactive element, through its decay, will irradiate the tumor in situ. Since radioactive rays travel some distance in tumors it is not necessary that every carcinoma cell bind the labeled antibody. The specificity of the monoclonal antibodies will permit a selective treatment of the tumor while avoiding the irradiation of innocent by-stander normal tissues, that could be dose limiting. Chemotherapy produces serious toxic effects on normal tissues, making the chemotherapy of carcinomas less than desirable, and the use of radiolabeled monoclonal antibodies a valid alterative.

Non-human antibodies raised against human epitopes have been used for the diagnosis and therapy of carcinomas as is known in the art. Also known are the methods for preparing both polyclonal and monoclonal antibodies. Examples of the latter are BrE-2, BrE-3 and KC-4 (e.g. U.S. Pat. Nos. 5,077,220; 5,075,219 and 4,708,930.

The KC-4 murine monoclonal antibody is specific to a unique antigenic determinant, the “antigen”, and selectivity binds strongly to neoplastic carcinoma cells and not to normal human tissue (U.S. Pat. No. 4,708,930 to Coulter). The antigen appears in two forms in carcinoma cells, only the smaller of these forms being expressed in the cell membrane. The larger form appears only in the cytoplasm and has an approximate 490 Kdalton molecular weight (range of 480,000-510,000). The second form occurs at a higher density of expression, is found both in the cytoplasm and the membrane of carcinoma cells and has an approximate 438 Kdalton molecular weight (range of 390,000-450,000) as determined by gel electrophoresis with marker proteins of known molecular weights. Labeled KC-4 was applied to the diagnosis and medical treatment of various carcinomas, particularly adenocarcinoma and squamous cell carcinoma regardless of the human organ site of origin.

The BrE-3 antibody (Peterson et al., Hybridoma 9:221 (1990); U.S. Pat. No 5,075,219) was shown to bind to the tandem repeat of the polypeptide core of human breast epithelial mucin. When the mucin is deglycosylated, the presence of more tandem repeat epitopes is exposed and the binding of the antibody increases. Thus, antibodies such as BrE-3 bind preferentially to neoplastic carcinoma tumors because these express an unglycosylated form of the breast epithelial mucin that is not expressed in normal epithelial tissue. The preferential binding combined with an observed low concentration of epitope for these antibodies in the circulation of carcinoma patients, such as breast cancer patients, makes antibodies having specificity for a mucin epitope highly effective for carcinoma radioimmunotherapy. A ⁹⁰Y-BrE-3 radioimmunoconjugate proved highly effective against human breast carcinomas transplated into nude mice. Human clinical studies showed the ⁹⁰Y-BrE-3 radioimmunoconjugate to considerably reduce the size of breast tumor metastases without any immediate toxic side effects. Moreover, an ¹¹¹In-BrE-3 radioimmunoconjugate was successfully used for imaging 15 breast cancer patients, providing excellent tumor targeting in 13 out of 15 of the patients. Out of all the breast tumor metastases occurring in another study, 86% were detected by ¹¹¹In-BrE-3. Unfortunately, 2 to 3 weeks after treatment, the patients developed a strong human anti-murine antibody (HAMA) response that prevented further administration of the radioimmunoconjugate. The HAMA response, which is observed for numerous murine monoclonal antibodies, precludes any long-term administration of murine antibodies to human patients. Similarly, other heterologous antibodies, when administered to humans, elicited similar antibody responses. The anti-heterologous human response is, thus, a substantial limiting factor hindering the successful use of heterologous monoclonal antibodies as therapeutic agents, which could, otherwise, specifically annihilate breast carcinomas, causing little or no damage to normal tissue and having no other toxic effects.

Chimeric antibodies are direct fusions between variable domains of one species and constant domains of another. Murine/human chimeric antibodies prepared from other types of B cells binding to other types of antigenic determinants have been shown to be less immunogenic in humans than wide murine antibodies. These proved to be less immunogenic but still in some cases an immune response is mounted to the rodent variable region framework region (FR). A further reduction of the “foreign” nature of the chimeric antibodies was achieved by grafting only the CDRs from a rodent monoclonial into a human supporting framework prior to its subsequent fusion with an appropriate constant domain (European Patent Application, Publication No. 239,400 to Winter; Riechmann, et al., Nature 332:323-327 (1988)). However, the procedures employed to accomplish CDR-grafting often result in imperfectly “humanized” antibodies. That is to say, the resultant antibody loses affinity (usually 2-3 fold, at best).

The ligand binding characteristics of an antibody combining site are determined primarily by the structure and relative disposition of the CDRs, although some neighboring residues also have been found to be involved in antigen binding (Davies, et al., Ann Rev. Biochem. 59:439-473 (1990)).

The technologies of molecular biology have further expanded the utility of many antibodies by allowing for the creation of class switched molecules whose functionality has been improved by the acquisition or loss of complement fixation. The size of the bioactive molecule may also be reduced so as to increase the tissue target availability of the antibody by either changing the class from the IgM to an IgG, or by removing most of the heavy and light chain constant regions to form an F_V antibody. Common to all of these potentially therapeutic forms of antibody are the required complementary determining regions (CDRs), which guide the molecule to its ligand, and the framework residues (FRs) which support the CDRs and dictate their disposition relative to one another. The crystallographic analysis of numerous antibody structures revealed that the antigen combining site is composed almost entirely of the CDR residues arranged in a limited number of loop motifs. The necessity of the CDRs to form these structures, combined with the appreciated hypervariability of their primary sequence, leads to a great diversity in the antigen combining site, but one which has a finite number of possibilities. Thus, its hypermutability and the limited primary sequence repertoire for each CDR would suggest that the CDRs derived for a given antigen from one species of animal would be the same derived from another species. Hence, they should be poorly immunogenic, if at all, when presented to a recipient organism.

Accordingly, there is still a need for a product of high affinity and/or specificity for carcinoma antigens suitable for the detection and therapy of carcinomas which elicits a lesser antibody response than whole non-human antibodies or chimeric antibodies containing, for instance the entire non-human variable region.

SUMMARY OF THE INVENTION

This invention relates to a humanized mouse monoclonal antibody and its glycosylated derivative which specifically and selectively bind to the human KC-4 antigen, the antibody consisting essentially of the variable regions of the light and heavy chains of the anti-KC-4 mouse antibody having the ATCC No. HB 8710 or HB 8709, wherein specific amino acids in the FR are substituted per chain with amino adds present in equivalent positions in antibodies of other species, and the constant region of a human antibody.

Also provided are the corresponding DNA and RNA segments encoding the monoclonal antibody, a hybrid vector carrying the DNA, and a transfected host thereof.

Still part of this invention are in vitro methods of diagnosing cancer and for conducting immunohistochemistry assays of tissue slices, an ex vivo method of purging neoplastic cells, and in vivo methods for imaging and therapy of cancer patients.

Other objects, advantages and features of the present invention will become apparent to those skilled in the art from the following discussion.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

This invention arose from a desire by the inventors to improve on antibody technology suitable for use in diagnostic and therapeutic applications, particularly for in vivo administration. The useful monoclonal antibodies obtained up to the present time have been prepared by fusing immortalized cell lines with B-cells of mouse or other animal origin. However, in general, heterologous antibodies may only be administered once to a human due to the detrimental immunological effects they elicit. This is true for most heterologous antibodies administered For example, the repeated administration of murine antibodies to a subject elicits a strong human anti-murine antibody (HAMA) response, which precludes their further utilization as therapeutic agents in humans. These heterologous antibodies initiate an immediate adverse reaction in many human patients and are, thus, rendered inaffective for further administration as therapeutic agents. On the other hand, human monoclonal hybridoma cell lines have not yet been very stable and have, therefore, not been suitable for the large scale, repeated production of monoclonal antibodies.

The present inventors, thus, have undertaken the preparation of anti-KC-4 humanized monoclonal antibodies maintaining the entire CDRs of the mouse antibodies of the same specificity and human constant regions, and substituting 7 amino acids in the heavy chain and 12 amino acids in the light chain, the substituted amino acids being positioned in the framework regions (FRs) and being selected from those present in equivalent positions in other human antibodies, and the constant regions of a human antibody. The

The PCR amplification conditions were as follows. All reagents as well as the GeneAmp PCR system 9600 were purchased from Perkin Elmer Cetus. Optimal PCR conditions were determined empirically for each pair of mutagenic primers. A matrix of conditions varying the concentration of MgCl₂, mutagenic primer, and template plasmid DNA were set up as follows. However, the annealing and extension temperatures during PCR may be varied.

2 μM primer JO59  150 nM each of primers JO61, 62, 63 and 64.
2 μM primer JO60  200 μM each of dGTP, dATP, TTP, and dCTP.
10 mM KCl         20 mM Tris-HCl pH 8.8
10 mM (NH4)2SO4   2 units per 100 μl reaction Vent DNA
0.1% Triton X-100 polymerase (New England Biolabs)
                  6 mM MgSO4

Example 20

Hot Start PCR for Humanization of anti-KC-4 Antibody

All the components of the PCR mixture, with the exception of Vent DNA polymerase, were mixed. The mixture was then dispensed in 19 μl aliquots in 5 PCR tubes. The reason for performing five independent reactions was to decrease the odds that unwanted mutations be isolated as a result of nucleotide misincorporation during PCR. The tubes were heated to 95° C. for 5 minutes and then cooled to 72° C. While at that temperature 1 μl of an appropriate Vent DNA polymerase dilution in 1×buffer was added to the reaction mixture (hot start). The temperature cycling then proceeds as follows.

[(96° C., 6 sec) (55° C., 10 sec) (72° C., 30 sec)] 3 cycles

[(96° C., 5 sec) (60° C., 10 sec) (72° C., 30 sec)] 29 cycles

72° C., 10 min

Example 21

Extra Final Extension for Humanized anti-KC-4 Antibody V_H DNA

After cycling, one extra final extension reaction was carried out Extra deoxyribonucleotide triphosphates (to 125 μM) and 1 unit of Vent DNA polymerase were added, and the mixture was heated to 72° C. for 10 minutes.

The resulting synthetic DNA fragment was digested with Dral and Nbel and inserted into the same restriction sites an intermediate plasmid construct encoding the corresponding murine heavy chain variable region as described in examples 23 to 25.

Example 22

Synthesis of anti-KC-4 Humanized Light Chain Variable Regions

The light chain variable region (V_L) genes were synthesized in a similar way as described in Examples 22 to 24 above for the heavy chain signal peptide variable regions. In this case, however, the complete signal peptide and the V_L encoding DNA were contained between EroRV and SaII. This DNA was inserted (ligated) into pBluescriptIIKs⁺ (Stratagene) as described in examples 23 and 25.

Example 23

Purification of Humanized anti-KC-4 PCR Products

The PCR products were then separated on a 0.8% agarose gel in 1XTAE buffer and 0.5 μg/ml ethidium bromide. The correct DNA bands were visualized with UV light (366 nm), excised from the gels and extracted with the GeneClean kit (Bio 101, La Jolla, Calif.).

Example 24

Ligation of Humanized anti-KC-4 DNA to Plasmids (Reclosure of Plasmid)

The ligation mixtures consisted of 5 μl extracted DNA, 2 μl 10× ligation buffer (NEB) 1 μl T4 DNA polymerase (NEB), 12 μl water. The amount of plasmid DNA may be varied depending of the intensity of the band extracted from the Gel. Ligations were carried out at room temperature for 2 hrs., or alteratively at 14° C. overnight.

Example 25

Transformation and Sequencing of Humanized anti-KC-4 DNA

The reclosed plasmids were then transformed into E. coli utilizing Inv alpha F′ competent cells purchased from Invitrogen Corporation, San Diego Calif. Plasmid DNA was then prepared from a few transformants and sequenced to verify that mutagenesis was successful.

Example 26

Hybrid Plasmid Preparation and Sequencing

Plasmid DNA was then prepared and sequenced to verify that the gene synthesis was successful. The anti-KC-4 humanized DNA sequences for the V_H and V_L segments are shown in Tables 21 and 22 below.

TABLE 21

Humanized anti-KC-4 Antibody VL DNA sequences

anti-KC-4 VL FR-HZ
ATG AAG TTG CCT GTT AGG CTG TTG GTG CTG ATG TTC TGG ATT CCT GCT
TCC AGC AGT GAT GTT TTG ATG ACC CAA ACT CCT CTC TCC CTG CCT GTC
ACT CCA GGA GAG CCA GCC TCC ATC TCT TGC AGA TCT AGT CAG AGC
ATT GTA CAT AGT AAT GGA AAC ACC TAT TTA GAA TGG TAC CTG CAG AAA
CCA GGC CAG TCT CCA CAG CTC CTG ATC TAC AAA GTT TCC ATC CGA TTT
TCT GGG GTC CCA GAC AGG TTC AGT GGC AGT GGA TCA GGG ACA GAT
TTC ACA CTC AAG ATC AGC AGA GTG GAG GCT GAG GAT GTG GGA ATT
TAT TAC TGC TTT CAA GGT TCA CAT GTT CCG TAC ACG TTC GGA GGG
GGG ACC AAG CTG GAA ATA AAA C (SEQ. ID. NO.: 48)

TABLE 22

Humanized anti-KC-4 Antibody VH DNA Sequences

anti-KC-4 VH FR-HZ
ATG GAC TTT GGG CTC AGC TTG GTT TTC CTT GTC CTT ATT TTA AAA GGT
GTC CAG TGT GAA GTG CAG ATG GTG GAG TCT GGG GGA GGC TTA GTG
CAG CCT GGA GGG TCC CTG AGA CTC TCC TGT GCA GCC TCT GGA TTC
GCT TTC AGT AGC TAT GCC ATG TCT TGG GTT CGC CAG GCT CCA GGG
AAG GGG CTG GAG TGG GTC GCA GAA ATT AGT AGT GGT GGT AAT TAC
GCC TAC TAT CAA GAC ACT GTG ACG GGC CGA TTC ACC ATC TCC AGA
GAC AAT TCC AAG AAC ACC CTG TAC CTG CAA ATG AAC AGT CTG AGG
GCT GAG GAC ACG GCC GTG TAT TAC TGT GCA AGG GAG GAC TAC GGT
ATC CCG GCC TGG TTT GCT TAC TGG GGC CAA GGG ACT CTG GTC ACT
GTC TCT AGT (SEQ. ID. No.: 49)

Example 27

Expression of anti-KC-4 Humanized Antibody

Two expression vectors pAG4622 and pAH4604 (Coloma, M. J., et al. (1992), supra) were used that were developed and provided by S. L. Morrison (Dept. of Microbiology and Molecular Genetics, UCLA). Any cDNA encoding a signal peptide and either the variable heavy chain or the variable light chain can, in principle, be inserted into these vectors resulting in a construction that encodes an IgG1, K, antibody with human constant regions. Synthetic DNA fragments were excised from their intermediate plasmids (see examples 21 and 22) with either EcoRV and Sal to be inserted into pAG4622(light chain vector), or with EcoRV and NbEl to be inserted into pAH4640 (heavy chain vector). The restriction and ligation reactions necessary to accomplish these operations were performed under the conditions stipulated by the enzyme manufacturers (New England Biolabs, Beverly, Mass.). Both the vectors and the inserts were purified from an agarose gel prior to ligation, using the Bio101 (La Jolla, Calif.) GeneClean kit (glass beads). The V_H and V_L regions in the final constructions were sequenced once again to verify that they were correct. The non-producer mycloma cell line SP2/0-Ag14, ATCC: CRL 1581, (Shulman M., et al. (1978), supra) was transfected with both plasmid constructions, and antibody producers were isolated following the recommendations outlined in Coloma et al. (Coloma, M. J. et al. (1992), supra) except that selection was done only for the uptake of hisD (by adding 5 mM histidinol to the medium and readjusting the pH to 7.4 with NaOH). Usually after ten days, stable transfectant colonies were established at a frequency of approximately 10⁻³ to 10⁻⁴. Colonies were then transferred to normal medium (without histidinol). The culture media were either Dulbeco's modified Eagle's medium (DME): fetal bovine serum (FBS), 90:10, v/v, or a mixture of DME:RPMI:FBS, 45:45:10, v/v/v. Penicillin and streptomycin were added to prevent bacterial growth.

The supernatants from stable transfectants were assayed for the presence of the antibodies. This was done by capturing the secreted chimeric antibody with a plate-bound goat anti-human-kappa chain antibody and developing with goat anti-human-gamma chain antibody, essentially as described previously (Coloma, M. J., (1992), supra) except that the secondary antibody was radiolabeled with ¹²⁵I. The supernatants were also assayed for binding to human milk fat globule (HMFG) as described previously (Ceriani R. L., et al., “Diagnostic Ability of Different Human Milk Fat Globule Antigens in Breast Cancer”, Breast Cancer Res. Treat. 15:161-174 (1990)). HMFG is bound to the microtiter plates as described previously (Ceriani, R. I. (1984), supra). Usually most colony supernatants were positive by both assays.

Colonies that secrete the highest level of antibody in the supernatants, as determined by these assays, were subcloned and subsequently adapted to serum-free medium for the purification of antibody. Serum free medium contains HL-1 supplement as directed by the manufacturer (Ventrex Labs, Portand, Me.).

Example 28

Half Humanized-Half Chimeric anti-KC-4 Antibody

An anti-KC-4 humanized light chain was paired with an anti-KC-4 non-humanized chimeric heavy chain by co-transfection of SP2/0Ag14 mycloma cells with hybrid plasmids carrying the respective DNA sequences and those of a human F_C. The resulting antibody was named “HuKC4V1” (ATCC No. HB 11454).

In addition, an anti-KC-4 humanized heavy chain was paired with an anti-KC-4 non-humanized chimeric light chain as described in Example 27 above. The resulting antibody was named “HuKC4V3” (ATCC No. HB 11456).

Example 29

Fully Humanized anti-KC-4 Antibody

An anti-KC-4 fully humanized antibody was prepared by pairing fully humanized anti-KC-4 light and heavy chains by co-transfection as described in Example 27 above. The fully humanized versions is named “HuKC4V2” (ATCC No. HB 11455).

Example 30

Determination of Affinity Constants for Fully Humanized anti-KC-4 Antibody

The secreted fully humanized antibody (HuKC4V2) was purified from culture supernatants using a Sepharose 4B -protein A column (Bio-Rad, Richmond, Calif.) as described by Ey et al. (Ey, P. L, et al. (1978), supra). Microtiter plates (Dynatech, Chantilly, Va.) were prepared as described by Ceriani et al. (Ceriani R. L., et al. (1992), supra) using successive layers of methylated BSA, glutaraldehyde, anti-β-galactosidase and the bacterial fusion protein 11-2 (a hybrid of β-galactosidase and human mammary mucin). Each well contained 388 mg of the 11-2 fusion protein. To each well were added 25 μl ¹²⁵I-KC-4 in RIA buffer (10% bovine calf serum, 0.3% triton X-100, 0.05% sodium azide pH 7.4, in phosphate buffer saline) and compete with 25 μl of either unlabeled murine or chimeric antibody in RIA buffer at the final concentrations of 130 pM, 850 pM, 1.3 nM, 4 nM, and 13 nM). Iodinations were performed with ¹²⁵I (17 Ci/mg, Nordion International). 50 μg anti-KC-4 monoclonal antibody (Coulter, Hialeah, Fla.) were labeled at a specific activity of 9.56 mCi/mg using the chloramine T method as described previously by Ceriani et al. (Ceriani R. L., et al., (198), supra).

The antibody-antigen affinity constants were determined by taking the reciprocal of the concentration of competing unlabeled monoclonal antibody that produced 50% binding as described by Sheldon et al. (Sheldon K., et al. (1987), supra).The protocol used to determine affinity constants was as described above except that in each case, an unlabeled antibody competed for binding to the antigen against the same radiolabeled antibody. The fully humanized antibody was shown to compete as well as anti-KC-4 murine antibody against radiolabeled anti-KC-4 murine antibody for binding to the KC-4G3 antigen.

Polyacrylamide gel electrophoresis was performed to insure that the antibody chains migrated as expected. The affinity binding constants of the murine, chimeric, half humanized and humanized antibodies were determined in independent competition assays. The binding affinities of the murine and anti-KC-4 and HuKC4V2 antibodies for the KC-4G3 antigen were determined to be similar.

Example 31

Histochemical Specificity of Fully Humanized Antibody

Immunohistochemical staining using the immunoperoxidase technique of consecutive human breast carcinoma tissue sections was used as a test to verify that the analogue antibodies retain the affinity for the KC-4G3 carcinoma antigen of the murine antibody. Breast carcinoma tissue sections were stained with the supernatant of the KC-4 murine and fully humanized transfected cells using the Vectastain ABC method (Vector Labs, Burlingame, Calif.). Both antibodies showed strong staining patterns.

The following Table 23 shows the results of the immunoperoxidase staining of five human breast carcinomas with either the standard anti-KC-4G3 murine or the fully humanized antibodies. Both stained the same tissues at a comparative level.

TABLE 23
Immunoperoxidase Staining of Human Breast
Carcinoma Tissue Sections with Murine and
Fully Humanized anti-KC-4 Antibodies
Breast Tumor	Murine Antibody	Fully Humanized Antibody
1	++	++
2	+++	+++
3	−	−
4	++	++
5	+++	+++

Example 32

Binding to HMFG of Half Humanized and Fully Humanized anti-KC-4 Antibodies

Tissue culture supernatants from transfections of all three anti-KC-4 variants of the humanized antibody were shown to bind the human milk fat globule (HMFG) as determined by radio-immunodetections.

Example 33

Half Humanized and Fully Humanized anti-KC-4 Antibodies Bind to Goat anti-Human κ or γ Antibodies

Tissue culture supernatants from transfections of all three variants of the anti-KC-4 humanized antibody were shown to bind in sandwich radioimmunodetections to both goat anti-human kappa chain antibody bound to microtiterplate wells (750 ng/well), and to radio-iodinated ¹²⁵I-labeled goat anti-human gamma chain antibodies.

The results of these sandwich assays demonstrate that both chains of the humanized antibodies indeed possess human kappa and gamma constant regions.

Example 34

Deduced Amino Acid Sequence of Humanized anti-KC-4 Variable Light and Heavy Chains

The amino acid sequences of the light and heavy chains of the analogue humanized antibody are shown in Tables 24 and 25 below. The actual amino acid sequences may be varied either to increase affinity for the antigen or to decrease immunogenicity in humans. Numerous variants of this sequence may be engineered in accordance with the invention.

TABLE 24
Humanized anti-KC-4 Antibody VL Analogue Sequence
anti-KC-4 VL FR-HZ
MKLPVRLLVL	MFWIPASSSD	VLMTQTPLSL	PVTPGEPASI	SCRSSQSIVH
SNGNTYLEWY	LQKPGQSPQL	LIYKVSIRFS	GVPDRFSGSG	SGTDFTLKIS
RVEAEDVGIY	YCFQGSHVPY	TFGGGTKLEI	K	(Seq. ID No: 50)

TABLE 25
Humanized anti-KC-4 Antibody VH Analogue sequence
anti-KC-4 VH FR-HZ
MDFGLSLVFL	VLILKGVQCE	VQMVESGGGL	VQPGGSLRLS	CAASGFAFSS
YAMSWVRQAP	GKGLEWVAEI	SSGGNYAYYQ	DTVTGRFTIS	RDNSKNTLYL
QMNSLRAEDT	AVYYCAREDY	GIPAWFAYWG	QGTLVTVSS	(Seq. ID No: 51)

Example 35

Hybridoma Cell Deposits

The following cell lines were deposited as present examples of the best mode of the invention. The hybridoma cell line expressing the anti-KC4 murine-human chimeric antibody was deposited with the ATCC on Nov. 13, 1992 under the Budapest Treaty, and has been assigned Accession No. HB 11201 (Chimeric anti-KC-4 1E8). The hybridoma cell lines expressing the anti-KC-4 fully humanized antibody (huKC4V2), and the half humanized anti-KC-4 antibodies (huKC4V1 and huKC4V3) were deposited with the ATCC on Sep. 23, 1993 and have been assigned Accession Nos. HB 11455 (Humanized HuKC-4 V2), HB 11454 (Half Humanized HuKC4V1), and HB 11456 (Half Humanized HuKC4V3).

The invention now being fully described, it will be apparent to one of ordinary skill in the art that many changes and modifications can be made thereto without departing from the spirit or scope of the invention as set forth herein.

Claims

1. A modified chimeric humanized anti-KC-4 antibody which selectively binds to the human KC-4 antigen, comprising (1) the humanized variable regions of the light and heavy chains of an anti-KC-4 murine antibody and (2) light and heavy chain constant regions of a human antibody, wherein the amino acid sequence of the variable regions comprises SEQ ID NOS: 50 and SEQ ID NO: 51.

2. The antibody of claim 1, wherein 12 amino acids in the light chain framework regions (FRs) are substituted and 7 amino acids in the heavy chain FRs are substituted.

3. The antibody of claim 2, wherein said amino acids in said framework regions are amino acids present in equivalent positions in antibodies of species other than murine.

4. An antibody according to claim 1, further comprising at least one glycosyl residue attached thereto.

5. A composition of matter, comprising the antibody of claim 1 and a carrier.

6. A cancer diagnostic kit, comprising the composition of claim 5, wherein said carrier is a pharmaceutically acceptable carrier for in vivo administration, and wherein said antibody is in radiolabeled form.

7. A cancer therapy kit, comprising the composition of claim 4 in a pharmaceutically-acceptable form, wherein said antibody is in radiolabeled form.

8. An in vitro cancer diagnostic kit, comprising the antibody of claim 1; and a solid support.

9. The in vitro kit of claim 8, further comprising one or more heterologous immunoglobulins that selectively bind to the constant region of the anti-KC-4 antibody.

10. The in vitro kit of claim 9, further comprising a label selected from the group consisting of a radioisotope, an enzyme, a phosphorescent molecule, and a fluorescent molecule, wherein said label is conjugated to said antibody of said immunoglobulin.

11. A hybridoma cell expressing the antibody of claim 1.

12. The hybridoma cell of claim 11, having the ATCC Accession No. HB 11455 (HuKC4V2).

13. A composition, comprising the hybridoma cell of claim 11, and a diluent or a carrier.

14. A method of determining the presence of cancer cells in a tissue or a blood sample thereof, comprising contacting a tissue or a blood sample with the anti-KC-4 antibody of claim 1 or an antigen binding fragment thereof, and detecting the presence of an immune complex formed between said anti-KC-4 antibody and the KC-4 antigen of said tissue or said blood sample thereof.

15. A method according to claim 14, wherein the tissue is excised from a subject prior to being contacted with the anti-KC-4 antibody, and wherein said contacting and detecting steps are carried out in vitro.

16. A method according to claim 15, further comprising comparing results obtained by analyzing said formed complexes with a standard cut-off value determined by comparing results obtained for normal and cancer patients' samples wherein a result above a cut-off value is indicative of cancer.

17. A method of imaging a cancer, comprising administering a pharmaceutically-acceptable composition comprising an effective imaging amount of the anti-KC-4 antibody of claim 1 or an antigen binding fragment thereof, in radiolabeled form, to a subject suspected of having a primary or metastasized cancer; and detecting any binding of the labeled antibody to cancer cells in the subject by external imaging.

18. A method of treating cancer in a subject, comprising administering to a subject in need of treatment an effective therapeutic amount of the antibody of claim 1, wherein the antibody is conjugated to a therapeutic agent.

19. A method of purging cancer cells from a biological sample, comprising contacting a biological sample obtained from a subject suspected of having cancer with the antibody of claim 1 and allowing said antibody to bind the KC-4 antigen associated with cancer cells in the sample to form an immune complex between said antibody and the antigen of said biological sample; and removing said immune complex from the biological sample.

20. A method according to claim 14, wherein said tissue sample is prepared for immunohistochemical analysis.

21. A method of purging cancer cells from a biological sample and replenishing the purged sample to a subject, comprising the method of claim 19, wherein said method further comprises reintroducing the purged biological sample to the subject.